This paper focuses on the statistical analysis of the energy fraction supplied by solar energy for solar water heating systems based on the f-chart method. An analysis is also presented for photovoltaic systems, where costs are linearly proportional to collector area. The uncertainty of the solar fraction is correlated with the monthly means of the global irradiation and the correlation coefficient between monthly means. Numerical examples for one location in Brazil and three locations in the United States are presented. These examples show that the uncertainty of the life cycle savings is significantly dependent on the uncertainty of the monthly means of the solar radiation data. The present analysis intends to provide a basic procedure that could be useful to make a straightforward feasibility analysis of a solar system. This is particularly interesting to evaluate the investment risk associated with photovoltaic plants, for which the capital costs are comparable to the advantages in saving electric energy from the utility grid under present scenarios in most places. (C) 2001 Elsevier Science Ltd. All rights reserved.